摘要： 提出一种新颖的两阶段决策树建树方法，在对数据集进行较粗的分类后，通过遗传 算法寻找规则集来建立决策树叶子节点。该方法可以同时对多个属性进行度量，并避免了决 策树的剪枝过程。

关键词: 数据挖掘, 决策树, 遗传算法

Abstract: This paper introduces a novel two-phrase-based decision tree constru ction method. It first classifies data set roughly, then uses genetic algorithm to construct rule set of the leaf nodes in tree. This method can measure multi a ttributes simultaneously, and avoid tree-pruning procedure.

中图分类号: 

• TP311.13